In recent years, Discontinuous Galerkin (DG) discretisations of non-linear 2D shallow water equations have evolved to be a viable tool for many coastal applications. As coastal flood simulations are time critical and computational resources are limited, there is a growing demand for the increase of computational efficiency and reduced run time. For that reason we have developed a DG storm surge model on a triangular and dynamically adaptive mesh. The dynamic mesh refinement and coarsening are driven by physics-based refinement indicators capturing major model sensitivities.

In this talk we will discuss source terms to be taken into account for storm surge simulations and validate their implementation using a set of numerical test cases. Furthermore, we demonstrate the model’s capability to reproduce known sensitivities of an idealised hurricane approaching a coast line. We will conclude with the definition and application of metrics that quantify the computational efficiency of the adaptive mesh refinement. This set of metrics will allow a novel insight into the comparison of uniform and adaptive simulations.

The presented work will include research conducted in collaboration with Jorn Behrens (University of Hamburg), Cristobal E. Castro (Universidad de Tarapaca), Frederic Dias (University College Dublin), and Stefan Vater (Free University of Berlin).

Acknowledgements: We kindly acknowledge support by the Irish Research Council (IRC) under the project "NIMBUS3: Next-Generation Integrated Model for Better and Unified Storm Surge Simulations” (GOIPD/2018/248).